rajunify123/freeswitch-mod-audio-fork repository overview

⁠FreeSWITCH with mod_audio_fork

A production-ready FreeSWITCH 1.10 Docker image pre-built with mod_audio_fork — a module that forks live call audio and streams it in real-time over a WebSocket connection to any remote server. Ideal for AI voice agents, speech recognition (ASR) pipelines, real-time transcription, and conversational IVR applications.

⁠What is mod_audio_fork?

mod_audio_fork attaches a media bug to a FreeSWITCH channel and streams raw L16 (16-bit PCM) audio over a WebSocket (ws:// or wss://) to your backend in real-time. Audio is never written to disk, making this ideal for privacy-sensitive or "no data at rest" deployments.

It also supports receiving audio back from the WebSocket server and playing it to the caller — enabling full-duplex AI-driven conversations.

⁠Quick Start

docker pull rajunify123/freeswitch-mod-audio-fork:v1

docker run -d \
  --name freeswitch \
  --network host \
  -e EXT_SIP_IP=<your-public-ip> \
  -e EXT_RTP_IP=<your-public-ip> \
  -v /path/to/your/conf:/usr/local/freeswitch/conf \
  rajunify123/freeswitch-mod-audio-fork:v1

⁠Exposed Ports

⁠Environment Variables

⁠Using mod_audio_fork

Once a call is established, trigger audio forking via the FreeSWITCH API:

uuid_audio_fork <uuid> start <wss-url> <mix-type> <sampling-rate> [metadata-json]

Parameters:

• uuid — FreeSWITCH channel UUID
• wss-url — WebSocket endpoint (ws:// or wss://)
• mix-type — mono | mixed | stereo
• sampling-rate — 8k or 16k
• metadata-json — Optional JSON sent as the first frame before audio begins

Example:

uuid_audio_fork 8cf9eb-... start wss://my-asr-server.com/stream stereo 8k {"caller":"12125551234"}

Stop streaming:

uuid_audio_fork <uuid> stop

⁠Audio Playback from WebSocket Server

The WebSocket server can send audio back to play to the caller:

{
  "type": "streamAudio",
  "data": {
    "audioDataType": "raw",
    "sampleRate": 8000,
    "audioData": "<base64-encoded-pcm>"
  }
}

⁠Events Fired by the Module

⁠Environment Variables for mod_audio_fork (build-time)

⁠Volumes

⁠Tags

⁠License

FreeSWITCH is licensed under the MPL 1.1⁠. mod_audio_fork is open-source (drachtio project).